Article

This Week in AI: November 13–19, 2026

AI news for November 13–19, 2026: GPT-6.2, Qwen4-Max, Gemini 4 Neon, Ignite, open models, Genie 4, Anthropic, and NVIDIA—with sources.

By Colin Michaels - Nov 22, 2026

Official Google artwork introducing Gemini 4 Neon

Trust & transparency

Evidence & disclosures

Evidence basis: Researched analysis. The article compares public evidence and does not claim hands-on testing unless a section says otherwise.

News roundup based on linked primary announcements and release documentation.

Sources checked: Nov 22, 2026

Linked references: 14 explicit sources in the article.

Synthetic media: Main series artwork is an AI-generated editorial illustration. Inline publisher artwork and video thumbnail are credited.

How evidence labels and corrections work

Friday’s roundup is arriving a little late, with a bigger model menu and new AI worlds to explore. Here’s what happened November 13–19, with demos and links to read more.

Coverage: November 13–19, 2026.

New Models

GPT-6.2 arrives in two flavors

November 17

OpenAI introduced GPT-6.2 Astra for demanding projects and Sol for everyday work with several steps. Both accept text and images through the API, and are rolling out in ChatGPT Work and Codex for eligible paid plans. OpenAI reports improvements in coding, computer use, and professional work; enterprise access depends on administrator settings.

Read OpenAI’s Astra and Sol announcement →

Qwen4-Max graduates from preview

November 16

Alibaba released Qwen4-Max after last week’s preview, with native vision and a million-token context window for lengthy coding, research, and work tasks. It is available through Qwen’s cloud services, API, Work, and Studio. Smaller Qwen4 models with downloadable weights are promised later; this announcement does not make Max an open-weight release.

Read Alibaba’s Qwen4-Max release →

Gemini 4 Neon takes on longer jobs

November 18

Google introduced Gemini 4 Neon for work that involves long documents, tools, and multiple steps. It is available through the Gemini API and AI Studio, with access also announced for Gemini for business and eligible paid Gemini and Antigravity users. Google positions Neon for longer, tool-heavy workflows, with performance comparisons reported in its announcement.

Explore Google’s Gemini 4 Neon announcement →

Official Google artwork introducing Gemini 4 Neon

Microsoft brings a model buffet to Ignite

November 17

Microsoft’s Ignite announcements span MAI Code-2 and Code-2-Flash, Cyber-2-Flash, Image-2.7, Realtime-1, and Decision-1.1, alongside Dragon Copilot transcription updates. Availability varies by model and product: the coding family reaches GitHub Copilot, while some Foundry access remains limited preview. Realtime-1 has an interest-registration route rather than blanket public availability.

See Microsoft’s Ignite model lineup →

Kling Image 4 aims for sharper details

November 18 announcement

Kling introduced Image 4, emphasizing native 8K images, more accurate text, and greater consistency when using reference images. It is available in Kling Creative Studio, with API access for developers and enterprises. Kling’s announcement pairs its quality claims with examples of detailed scenes, readable lettering, and reference-based creations.

See Kling’s Image 4 announcement and examples →

Rumble 2 Flash opens up speech transcription

November 18

Reson8 released Rumble 2 Flash, a 0.5-billion-parameter speech-recognition model covering more than 40 languages. Its downloadable weights use Apache 2.0 licensing, and it supports automatic language detection and controls for details such as numbers and filler words. The release includes Transformers and vLLM support for developers building transcription tools.

Read Reson8’s Rumble 2 Flash release →

Technology Advances

Sage brings molecules and materials into the conversation

Repository released November 15–16

InternLM published Sage-S2.1-Flash, a scientific multimodal model built to work with inputs such as molecules, crystals, and proteins. Its tools support retrieval, code execution, and checking during longer scientific tasks. Model weights and code are available under MIT licensing. The release focuses on scientific reasoning across these different kinds of data.

Explore the official Sage model card and downloads →

Genie 4 opens its doors a little wider

November 18–19

Google expanded Genie 4 access through experimental Project Genie for eligible Ultra subscribers, then announced access through the Gemini API and AI Studio. The world model generates interactive environments with synchronized audio and responds to movement and live prompts. The update moves October’s model announcement into more people’s hands.

Read about Genie 4 in Project Genie →

Watch the companion video on YouTube

Watch: Introducing the new Project Genie, powered by Genie 4 — Google DeepMind; official launch demo. →

Google DeepMind’s official Project Genie video thumbnail

Read the Genie 4 API announcement →

More academic researchers can apply for ChatGPT access

November 16; applications open November 23

OpenAI announced another round of its academic researcher program: 50,000 grants, each covering a researcher and up to two collaborators with ChatGPT Plus access. Applications open November 23, subject to eligibility and availability. The new round is part of the program’s broader goal of reaching 200,000 researchers.

Read the academic researcher program details →

Companies Making News

Anthropic prices its public offering

November 18–19

Anthropic priced its initial public offering at $83 a share, with trading under ANT beginning on Nasdaq the following day. The company’s announcement set out the offering terms, while AP covered the market debut. It marks a concrete next step after last week’s IPO preparations.

Read Anthropic’s IPO pricing announcement →

Read AP’s report on the trading debut →

NVIDIA creates a board oversight committee

November 17

NVIDIA’s board established a standing National Security and Export Control Oversight Committee, chaired by Suzanne Nora Johnson. The committee will oversee related company policies and receive reporting from NVIDIA’s security and global trade teams. It is a specific governance development following the leadership news covered last week. The new group will also draw on independent outside expertise.

Read NVIDIA’s committee announcement →